A VC Says the ‘All Your Eggs in the AI Basket' Trade Is Finally Cracking — Here's Where the Money Goes Next

AI-heavy mega caps are under pressure as investors rotate away from hyperscalers, with the NASDAQ 100 down 3.27% on the week and Apple, Alphabet, and Microsoft all lower amid capex concerns. Microsoft spent $30.88 billion on Q3 capex, Alphabet guided 2026 capex to $175 billion-$185 billion, while suppliers such as Micron and Taiwan Semiconductor are benefiting from the reallocation. Micron beat with $41.456 billion in revenue versus expectations and raised Q4 guidance, reinforcing the view that markets are rewarding the companies building AI infrastructure rather than the biggest AI spenders.

Analysis

This is less a broad “AI selloff” than a forced re-rating of the value chain: the market is starting to discriminate between capex-intense demand creators and the second-order beneficiaries that monetize the same buildout with less balance-sheet drag. If hyperscalers are proving they can still grow, the problem is not top-line demand but incremental return on each dollar of compute spend; that tends to compress multiples first in the buyers and then, with a lag, reprice the suppliers once order timing normalizes. The near-term winners are memory and foundry exposure because their revenue is now tied to scarcity and mix, not just unit volume.

The key risk is that this rotation can overshoot into a quality unwind if investors conclude AI capex is becoming self-cannibalizing. That would be most dangerous for names with the longest-duration earnings streams and the highest implied growth assumptions, because even a small change in terminal multiple overwhelms current EPS beats. The time horizon matters: over days to weeks, positioning and narrative can keep penalizing the large platforms; over months, the market will likely refocus on whether capex converts into paid usage, margin expansion, and backlog quality.

Contrarianly, the move in Alphabet looks more like a setup than a verdict. It has the clearest combination of a lower starting valuation, multiple growth engines, and enough balance-sheet flexibility to absorb another year of heavy investment without threatening capital returns. The consensus is still pricing the spend as pure dilution, but if even a portion of that capex improves cloud attach, model tooling, or autonomous optionality, the market will have to re-rate the stock before the next earnings cycle. Apple is the weaker dip-buy because capital returns can cushion downside, but they do not solve the market’s concern that its AI story remains more embedded than monetized.

Key Decisions for Investors

• Add GOOGL on weakness over the next 1-3 sessions; target a 3-6 month mean reversion trade with a favorable 15-20% upside if valuation closes part of the gap to peers, with downside bounded by strong cash generation and buyback support.
• Stay long TSM as the cleanest AI plumbing proxy for 3-6 months; the risk/reward remains attractive if hyperscaler capex merely slows rather than reverses, but trim if order commentary suggests digestion beyond one quarter.
• Use MSFT weakness to express a relative-value short vs GOOGL rather than outright short exposure; the risk is that Microsoft’s revenue quality and enterprise lock-in reassert faster than sentiment expects, so keep this as a tactical 1-2 quarter pair.
• Avoid chasing AAPL as a primary AI beneficiary; if used at all, treat it as a defensive capital-return name and prefer selling elevated volatility or waiting for a deeper drawdown before adding.
• For higher conviction trading, consider a short basket of hyperscaler beta against long MU/TSM as a pair trade for the next earnings window; the thesis is that capex intensity remains rewarded in suppliers before it is rewarded in platform multiples.